Depression affects a large portion of the world's population. Though treatable, two thirds of patients will not respond sufficiently to two or more standard pharmacotherapies and will be defined as treatment resistant (TRD). Quality of life for these individuals is extremely low and unremitting symptoms lead to loss of productivity, impaired social relationships, high health care costs, and in some cases, loss of life by suicide. Though several different brain networks are implicated, despite much research, the mechanisms causal to depression and its successful treatment remain unclear. The overarching goal of the current proposal is to leverage optimized non-invasive MRI technologies and normative data available through the NIMH/NIA-funded Human Connectome Project (HCP, U54 MH091657) to 1) identify connectome-specific correlates and predictors of successful treatment outcome to 3 therapeutic interventions, each with a rapid onset of action and to 2) characterize alterations in neural connectivity associated with individual clinical, behavioral and physiologica differences across TRD. Following harmonization of HCP MRI protocols, structural, functional and diffusion MRI data and behavioral testing batteries modeled from the HCP Lifespan protocol with added clinical assessments will be collected. Arterial spin labeling (ASL) perfusion MRI, measuring cerebral blood flow, and peripheral blood measures of gene function will supplement these protocols. Our first aim is longitudinal and will determine whether changes in brain network connectivity relate to and predict response to fast-acting perturbations known to elicit robust antidepressant effects. These perturbations include electroconvulsive therapy (ECT), serial ketamine infusion and total sleep deprivation (TSD). Since TRD includes different categorical diagnoses such as unipolar and bipolar depression and other comorbidities, our second specific aim is cross-sectional and will determine if heterogeneity in behavioral and symptom profiles, clinical histories and sex and age contribute to variations in the patterns of altered structural and functional connectivity in TRD. Subjects will include 200 patients clinicall eligible to receive ECT (n=60), serial ketamine (n=60) or TSD (n=80) and 140 controls, combining control data collected locally (n=40) with control data from the HCP resource (n=100). Each patient will receive MRI, behavioral/cognitive testing and a blood draw before and after completing one of the interventions. Behavioral constructs and sub-constructs of interest will include cognitive control, negativity bias and rumination and reward hypersensitivity, widely implicated in depression, functions that are governed by prefrontal and anterior cingulate cortex (cognitive control, mood regulation) and amygdala, hippocampus, ventral striatum/pallidum (emotion and reward) regions and circuitry. Data will be released to the scientific community through the Connectome Coordination Facility. The infrastructure of the HCP provides an unprecedented opportunity for to discover the mechanisms of disease and treatment response, which could lead to more effective treatment strategies based on individual connectivity profiles.